Seedlings from a northern and a southern provenance of black spruce (Picea mariana Mill. BSP) from eastern Canada were exposed to 37 or 71 Pa of carbon dioxide (CO2) during growth, cold hardening and dehardening in a greenhouse. Bud phenology, cold tolerance and photosynthetic efficiency were assessed during the growing and over-wintering periods. Bud set occurred earlier in elevated [CO2] than in ambient [CO2], but it was later in the southern provenance than in the northern provenance. An increase in seedling cold tolerance in early fall was related to early bud set in elevated [CO2]. Maximal photosystem II (PSII) photochemical efficiency (Fv/Fm), effective quantum yield (ΦPSII), photochemical quenching (qP), light-saturated photosynthesis (Amax), apparent quantum efficiency (α'), light-saturated rate of carboxylation (Vcmax) and electron transport (Jmax) decreased during hardening and recovered during dehardening. Although Amax and α' were higher in elevated [CO2] when measured at the growth [CO2], down-regulation of photosynthesis occurred in elevated [CO2] as shown by lower Fv/Fm, ΦPSII, Vcmax and Jmax. Elevated [CO2] reduced gene expression of the small subunit of Rubisco and also decreased chlorophyll a/chlorophyll b ratio and nitrogen concentration in needles, confirming our observation of down-regulation of photosynthesis. Elevated [CO2] increased the CO2 diffusion gradient and decreased photorespiration, which may have contributed to enhance Amax despite down-regulation of photosynthesis. Total seedling dry mass was higher in elevated [CO2] than in ambient [CO2] at the end of the growing season. However, because of earlier bud formation and cold hardening, and down-regulation of photosynthesis during fall and winter in elevated [CO2], the treatment difference in dry mass increment was less by the end of the winter than during the growing season. Differences in photosynthetic rate observed during fall, winter and spring account for the inter-annual variations in carbon assimilation of black spruce seedlings: our results demonstrate that these variations need to be considered in carbon budget studies.